64-bit MATLAB on a Mac (options and suggestions for the future)

I know that Psychophysics Toolbox may not work on the Mac OS Lion and does not work within 64 bit MATLAB on a Mac.

Is there an estimate regarding when and if PTB will work on 64 bit MATLAB on a Mac?

I'm asking in part to have a contingency plan for my own lab. And I'm curious what others labs are doing. Our university has a MATLAB site license and the license key will only work on 32 bit versions for another year or so, after which we will all need to shift to the 64 bit version.

I'm also asking because I'm planning to offer a new course in the fall that will cover core programming and mathematical concepts for graduate students in psychology and neuroscience. I've thought to design the course in a way that students can do MATLAB (with PTB) or Python (e.g., with PsychPy). If PTB will no longer be an option for some of our graduate students, then I will definitely design the course that way from the get-go.

Does everything that works on 32-bit MATLAB on a Mac work also on Octave on a Mac? When I looked at Octave years ago, it's libraries were incomplete. I assume they're better now.

If so, that's certainly an option. I've encouraged people to consider doing things in Python if only so they don't feel like they need to spend 10's of thousands of dollars on MATLAB licenses to set up a lab after they leave. If Octave is an option, and if PTB works on versions of Octave that run on Mac (OS Lion included) that would be great.

Vanderbilt has a $100,000 a year site license with Mathworks. They agreed to let the licenses work on the 32-bit versions for a year or two. But no longer.

From: "Allen W. Ingling" <allen.ingling@...>
Date: Mon, 5 Mar 2012 11:34:00 -0600
To: "psychtoolbox@yahoogroups.com" <psychtoolbox@yahoogroups.com>
Cc: "Palmeri, Thomas J" <tom.palmeri@...>, "Thomas G. Fikes" <fikes@...>
Subject: Re: [psychtoolbox] 64-bit MATLAB on a Mac (options and suggestions for the future)



For those confronting the looming 32-bit MATLAB expiration, why is Octave not a suitable alternative to MATLAB for your purposes?  Is it the lack of good Octave IDE on OS X? Are there other issues with Octave?  

- Allen




On Mar 5, 2012, at 11:12 AM, Thomas G. Fikes wrote:

Thomas (and others; cc: Dan at Mathworks):


I'm facing a similar dilemma here with the MATLAB programming class I teach for psych and neuroscience undergrads. Although our classroom and research licenses appear to allow for installing any previous version of MATLAB (including 32 bit versions -- we're running R2010a), I was surprised to find out this year that the student version available to my students is ONLY the current and next most recent version... which means they can't get a copy of MATLAB that will run PTB. Although I look forward to 64 bit PTB, I know what kind of development time is involved for that. In the mean time, perhaps we can put pressure on Mathworks to support older versions for our students (and apparently for Thomas's classroom license as well).

I've certainly registered my complaint with Mathworks (the customer service rep I talked with seemed particularly unconcerned -- "That's just the way we do it"). If anyone has any pull there, or just wants to join in registering a voice for MATLAB support on the mac, please join me.

~~~~~~~~~~~~~~~~~~~~~~~
Thomas G. Fikes
Department of Psychology
Westmont College
Santa Barbara, CA 93108
~~~~~~~~~~~~~~~~~~~~~~~



On Mar 5, 2012, at 8:49 AM, thomas.palmeri wrote:

 

I know that Psychophysics Toolbox may not work on the Mac OS Lion and does not work within 64 bit MATLAB on a Mac.

Is there an estimate regarding when and if PTB will work on 64 bit MATLAB on a Mac?

I'm asking in part to have a contingency plan for my own lab. And I'm curious what others labs are doing. Our university has a MATLAB site license and the license key will only work on 32 bit versions for another year or so, after which we will all need to shift to the 64 bit version.

I'm also asking because I'm planning to offer a new course in the fall that will cover core programming and mathematical concepts for graduate students in psychology and neuroscience. I've thought to design the course in a way that students can do MATLAB (with PTB) or Python (e.g., with PsychPy). If PTB will no longer be an option for some of our graduate students, then I will definitely design the course that way from the get-go.

Hi tom,
I don't know if anything has changed in the meantime, but I'm pretty sure that the only supported Octave version is 3.2.4 for PTB right now...
Best,
Dee

On Tue, Mar 6, 2012 at 14:03, Thomas G. Fikes <fikes@...> wrote:

Hi Allen & others:

To be honest, I just haven't invested the time to give Octave a real try. Lack of an IDE is an issue, but given your prompt and Mario's description, it sounds like it's worth it for running PTB experiments... even if I have to go back to MATLAB with my data to use all the signal processing and graphics mfiles I've invested so much time in.

I'm looking for an experience that will be positive for my students, which includes an install that doesn't require a PhD in computer science. I downloaded Gnu Octave 3.4.0 binary today. Very nice (& other than some misleading commands on the web site, not a bad install). But the Octave that launches is 64 bit. Some web sources suggest that the universal binary also includes a 32 bit version, but an hour or so searching the forums hasn't led me to a solution on how to get it to launch in 32 bit mode.

Perhaps I'll have to manually build a 32-bit version, but if anyone knows how to get the universal binary to run in 32 bit mode I'd be grateful.

I'm running Mac OS 10.6.8. At launch in the terminal I see 'Octave was configured for "x86_64-apple-darwin10.7.3".' and the ver command gives this info
----------------------------------------------------------
GNU Octave Version 3.4.0
GNU Octave License: GNU General Public License
Operating System: Darwin 10.8.0 Darwin Kernel Version 10.8.0: Tue Jun 7 16:33:36 PDT 2011; root:xnu-1504.15.3~1/RELEASE_I386 i386
----------------------------------------------------------

Thanks,
-tom



On Mar 5, 2012, at 9:34 AM, Allen W. Ingling wrote:

>
>
> For those confronting the looming 32-bit MATLAB expiration, why is Octave not a suitable alternative to MATLAB for your purposes? Is it the lack of good Octave IDE on OS X? Are there other issues with Octave?
>
> - Allen
>
>
>


If we all chipped in a bit from our grants, we each wouldn't need to pay very much, I think.  I'm willing to contribute.  We could even hire someone just to do the 64-bit port, though I don't know how much that would cost...

keith

On Mar 8, 2012, at 12:44 PM, IanA wrote:

 

--- In psychtoolbox@yahoogroups.com, "Mario" <mario.kleiner@...> wrote:

> Wrt. to a 64-Bit version for os/x i can't give you reliable estimates, just that most of it will very likely complete for the most part within this year. I'm still the only person doing any serious amount of C development on the toolbox and PTB development is still just an insanely time intensive "hobby" for a large part, beyond the bits i actually need for my own projects. I need to get a thesis finished and submitted. Once that is done i'll start to look into the 64-bit port. Now the thesis will be done, according to some outdated estimates of mine, in -6 months, -3 months, now and in approximately 3 months...

Dear All,

There are a substantial number of labs using PTB on OS X. Does anyone know of a viable route to get some focused funding to allow Mario the resources to do this. It would be great to see the PTB move forward with something other than the singular sterling efforts of Mario, or at least with some funding to purchase him a modern computer and Matlab licence to more easily fix things.

Ian


I'm interested in chipping in, too! I don't have a suggestion in term of how but if a system is set up, I'm more than willing to help!

--- In psychtoolbox@yahoogroups.com, Keith Schneider <keiths@...> wrote:
>
> If we all chipped in a bit from our grants, we each wouldn't need to pay very much, I think. I'm willing to contribute. We could even hire someone just to do the 64-bit port, though I don't know how much that would cost...
>
> keith
>
> On Mar 8, 2012, at 12:44 PM, IanA wrote:
>
> > --- In psychtoolbox@yahoogroups.com, "Mario" <mario.kleiner@> wrote:
> >
> > > Wrt. to a 64-Bit version for os/x i can't give you reliable estimates, just that most of it will very likely complete for the most part within this year. I'm still the only person doing any serious amount of C development on the toolbox and PTB development is still just an insanely time intensive "hobby" for a large part, beyond the bits i actually need for my own projects. I need to get a thesis finished and submitted. Once that is done i'll start to look into the 64-bit port. Now the thesis will be done, according to some outdated estimates of mine, in -6 months, -3 months, now and in approximately 3 months...
> >
> > Dear All,
> >
> > There are a substantial number of labs using PTB on OS X. Does anyone know of a viable route to get some focused funding to allow Mario the resources to do this. It would be great to see the PTB move forward with something other than the singular sterling efforts of Mario, or at least with some funding to purchase him a modern computer and Matlab licence to more easily fix things.
> >
> > Ian
> >
> >
>
--- In psychtoolbox@yahoogroups.com, "mlrogers2002" <mlrogers2002@...> wrote:
>
>
> There are lots of things I'd like to see happen with PTB, and I'd love to work on it, but... I don't have the time to do it alone. I could put in time if others who are also skilled at programming and science (and electronics is good too), or just have a need for it, wanted to organize something together - maybe we need a foundation of sorts?
>

Care to share what those "lots of things" are? We have a feature request page on the Wiki, but i don't see your name there. Another new good place for feature requests now is the Google issue tracker: <http://code.google.com/p/psychtoolbox-3/issues/list>

If you tell people what you want, maybe they point you to existing solutions or self-organize? Or if you just start contributing code, maybe someone else finds it useful and wants to work with you?

My experience with code contributions and volunteers in the last 6 years was that the amount of actual useful work contributed was inversely proportional to the amount of setup and startup overhead they required before they would actually do anything. Some people with phd's in computer-science and other credentials for coding skills and some companies even managed to delay development of important new features by over a year by "offering help" and requiring setup overhead, without ever doing anything substantial in the end.

For small code contributions people usually just e-mail me updated files or preferrably patches for review and integration. People who prove to be consistent/competent providers of improvements usually quickly get commit access to our repo. And since our main development repo is on GitHub, based on the git version control system... <https://github.com/kleinerm/Psychtoolbox-3>

... it should rather simple to just fork that repo, do complex development and then send pull requests, or coordinate collaborative development of more complex or intrusive features. In the end you'd need to coordinate with me anyway for complex stuff so there won't be clashes in the development of features.

I don't see how a foundation could help with software development? Also, really large scale software projects like Linux were able to grow for a long time without a foundation behind them. Imho it would just increase the administrative overhead without increasing the available manpower. This btw. is the result of the election of the board of directors of the XOrg foundation:

<http://www.phoronix.com/scan.php?page=news_item&px=MTA3MTc>

Despite providing the bases of pretty much any operating systems GUI (except MS-Windows) since 25 yeras, and having hundreds of active developers from probably dozens of companies (e.g. Intel, IBM, AMD, VMWare, RedHat, Oracle, ...), academia and volunteers, and 144 official members, they barely managed to get enough voters to complete elections. I also know that there is quite a bit of paperwork involved, because i attended their 2010 meeting (Video here: <http://www.youtube.com/watch?v=peCG_aM45-o>).

I assume a foundation would be useful if there were significant donations of money to the project involved, significant enough that people don't trust or can't use a less formal process of handling donations.

Wrt. to funding of ptb development, i don't know. It is nice to know that at least a few people would be willing to donate. We've toyed half-seriously in the past with something like a PayPal button or similar, but never thought it through.

In the past (many years ago), i think Denis Pelli and NYU had some "Core vision grant" with some money allocated to ptb's development, but i don't know the details. I guess if we tried to do something similar in the future we would need to proof ptb's wide-spread usage and usefulness by something like a large citation count on a to-be-published article about ptb-3, or having some list of grant-supported research projects that use ptb-3. For ptb-2 such lists were collected, <http://psychtoolbox.org/PTB-3/grants.html>, but not updated in at least 5 years afaik. Maybe we should revive that.

Our download counter has exceeded 88888 downloads and the conservative count of unique installations also just passed the 30000 mark last week, but i doubt that would really count as proof, given that these counts are created by ourselves, on servers under our control, anonymously without a way to track back who installed ptb where.

Last week i stumbled over this interesting approach from the world of startups: Crowd-funding of scientific research projects: <http://www.petridish.org/>, article about it here: <http://techcrunch.com/2012/03/08/petridish-aims-to-crowdfund-science-and-research-projects/>. Doing a 64-bit os/x port of ptb would sound rather boring though, compared to the featured projects.

At the moment, ptb's development is not officially formally funded at all. Most of my work was just spare-time coding on late evenings, weekends and "vacations". Since ptb also became a core part of my regular PhD project during the last years, i was also spending some fraction of my graduate student work time on it, following the "eat your own dog food" approach to software testing, while being supported by a stipend from the Max Planck Society, specifically by Prof. Dr. Heinrich Buelthoff at the MPI for Biological Cybernetics in Tuebingen. So currently the development is basically powered by Prof. Buelthoff's goodwill (thank you!) and my obsessive compulsive approach to "recreational coding" and geeky definition of fun. Then there are the occassional donations of hardware for development and testing, e.g., from Cambridge Research Systems and VPixx Inc., even a FireGL graphics card from AMD and other bits...

Not sure how to conclude this, but there's more work to do before the beta release.
-mario



> --- In psychtoolbox@yahoogroups.com, "ptballthumbs" <hyunah1112@> wrote:
> >
> > I'm interested in chipping in, too! I don't have a suggestion in term of how but if a system is set up, I'm more than willing to help!
> >
> > --- In psychtoolbox@yahoogroups.com, Keith Schneider <keiths@> wrote:
> > >
> > > If we all chipped in a bit from our grants, we each wouldn't need to pay very much, I think. I'm willing to contribute. We could even hire someone just to do the 64-bit port, though I don't know how much that would cost...
> > >
> > > keith
> > >
> > > On Mar 8, 2012, at 12:44 PM, IanA wrote:
> > >
> > > > --- In psychtoolbox@yahoogroups.com, "Mario" <mario.kleiner@> wrote:
> > > >
> > > > > Wrt. to a 64-Bit version for os/x i can't give you reliable estimates, just that most of it will very likely complete for the most part within this year. I'm still the only person doing any serious amount of C development on the toolbox and PTB development is still just an insanely time intensive "hobby" for a large part, beyond the bits i actually need for my own projects. I need to get a thesis finished and submitted. Once that is done i'll start to look into the 64-bit port. Now the thesis will be done, according to some outdated estimates of mine, in -6 months, -3 months, now and in approximately 3 months...
> > > >
> > > > Dear All,
> > > >
> > > > There are a substantial number of labs using PTB on OS X. Does anyone know of a viable route to get some focused funding to allow Mario the resources to do this. It would be great to see the PTB move forward with something other than the singular sterling efforts of Mario, or at least with some funding to purchase him a modern computer and Matlab licence to more easily fix things.
> > > >
> > > > Ian
> > > >
> > > >
> > >
> >
>
Hi Mario, I thought it would be a good idea for a bunch of us users to get together and pay *you*.  I didn't realize that you weren't being paid at all.

keith

On Mar 18, 2012, at 4:59 PM, Mario wrote:

 



--- In psychtoolbox@yahoogroups.com, "mlrogers2002" <mlrogers2002@...> wrote:
>
>
> There are lots of things I'd like to see happen with PTB, and I'd love to work on it, but... I don't have the time to do it alone. I could put in time if others who are also skilled at programming and science (and electronics is good too), or just have a need for it, wanted to organize something together - maybe we need a foundation of sorts?
>

Care to share what those "lots of things" are? We have a feature request page on the Wiki, but i don't see your name there. Another new good place for feature requests now is the Google issue tracker: <http://code.google.com/p/psychtoolbox-3/issues/list>

If you tell people what you want, maybe they point you to existing solutions or self-organize? Or if you just start contributing code, maybe someone else finds it useful and wants to work with you?

My experience with code contributions and volunteers in the last 6 years was that the amount of actual useful work contributed was inversely proportional to the amount of setup and startup overhead they required before they would actually do anything. Some people with phd's in computer-science and other credentials for coding skills and some companies even managed to delay development of important new features by over a year by "offering help" and requiring setup overhead, without ever doing anything substantial in the end.

For small code contributions people usually just e-mail me updated files or preferrably patches for review and integration. People who prove to be consistent/competent providers of improvements usually quickly get commit access to our repo. And since our main development repo is on GitHub, based on the git version control system... <https://github.com/kleinerm/Psychtoolbox-3>

... it should rather simple to just fork that repo, do complex development and then send pull requests, or coordinate collaborative development of more complex or intrusive features. In the end you'd need to coordinate with me anyway for complex stuff so there won't be clashes in the development of features.

I don't see how a foundation could help with software development? Also, really large scale software projects like Linux were able to grow for a long time without a foundation behind them. Imho it would just increase the administrative overhead without increasing the available manpower. This btw. is the result of the election of the board of directors of the XOrg foundation:

<http://www.phoronix.com/scan.php?page=news_item&px=MTA3MTc>

Despite providing the bases of pretty much any operating systems GUI (except MS-Windows) since 25 yeras, and having hundreds of active developers from probably dozens of companies (e.g. Intel, IBM, AMD, VMWare, RedHat, Oracle, ...), academia and volunteers, and 144 official members, they barely managed to get enough voters to complete elections. I also know that there is quite a bit of paperwork involved, because i attended their 2010 meeting (Video here: <http://www.youtube.com/watch?v=peCG_aM45-o>).

I assume a foundation would be useful if there were significant donations of money to the project involved, significant enough that people don't trust or can't use a less formal process of handling donations.

Wrt. to funding of ptb development, i don't know. It is nice to know that at least a few people would be willing to donate. We've toyed half-seriously in the past with something like a PayPal button or similar, but never thought it through.

In the past (many years ago), i think Denis Pelli and NYU had some "Core vision grant" with some money allocated to ptb's development, but i don't know the details. I guess if we tried to do something similar in the future we would need to proof ptb's wide-spread usage and usefulness by something like a large citation count on a to-be-published article about ptb-3, or having some list of grant-supported research projects that use ptb-3. For ptb-2 such lists were collected, <http://psychtoolbox.org/PTB-3/grants.html>, but not updated in at least 5 years afaik. Maybe we should revive that.

Our download counter has exceeded 88888 downloads and the conservative count of unique installations also just passed the 30000 mark last week, but i doubt that would really count as proof, given that these counts are created by ourselves, on servers under our control, anonymously without a way to track back who installed ptb where.

Last week i stumbled over this interesting approach from the world of startups: Crowd-funding of scientific research projects: <http://www.petridish.org/>, article about it here: <http://techcrunch.com/2012/03/08/petridish-aims-to-crowdfund-science-and-research-projects/>. Doing a 64-bit os/x port of ptb would sound rather boring though, compared to the featured projects.

At the moment, ptb's development is not officially formally funded at all. Most of my work was just spare-time coding on late evenings, weekends and "vacations". Since ptb also became a core part of my regular PhD project during the last years, i was also spending some fraction of my graduate student work time on it, following the "eat your own dog food" approach to software testing, while being supported by a stipend from the Max Planck Society, specifically by Prof. Dr. Heinrich Buelthoff at the MPI for Biological Cybernetics in Tuebingen. So currently the development is basically powered by Prof. Buelthoff's goodwill (thank you!) and my obsessive compulsive approach to "recreational coding" and geeky definition of fun. Then there are the occassional donations of hardware for development and testing, e.g., from Cambridge Research Systems and VPixx Inc., even a FireGL graphics card from AMD and other bits...

Not sure how to conclude this, but there's more work to do before the beta release.
-mario

> --- In psychtoolbox@yahoogroups.com, "ptballthumbs" <hyunah1112@> wrote:
> >
> > I'm interested in chipping in, too! I don't have a suggestion in term of how but if a system is set up, I'm more than willing to help!
> >
> > --- In psychtoolbox@yahoogroups.com, Keith Schneider <keiths@> wrote:
> > >
> > > If we all chipped in a bit from our grants, we each wouldn't need to pay very much, I think. I'm willing to contribute. We could even hire someone just to do the 64-bit port, though I don't know how much that would cost...
> > >
> > > keith
> > >
> > > On Mar 8, 2012, at 12:44 PM, IanA wrote:
> > >
> > > > --- In psychtoolbox@yahoogroups.com, "Mario" <mario.kleiner@> wrote:
> > > >
> > > > > Wrt. to a 64-Bit version for os/x i can't give you reliable estimates, just that most of it will very likely complete for the most part within this year. I'm still the only person doing any serious amount of C development on the toolbox and PTB development is still just an insanely time intensive "hobby" for a large part, beyond the bits i actually need for my own projects. I need to get a thesis finished and submitted. Once that is done i'll start to look into the 64-bit port. Now the thesis will be done, according to some outdated estimates of mine, in -6 months, -3 months, now and in approximately 3 months...
> > > >
> > > > Dear All,
> > > >
> > > > There are a substantial number of labs using PTB on OS X. Does anyone know of a viable route to get some focused funding to allow Mario the resources to do this. It would be great to see the PTB move forward with something other than the singular sterling efforts of Mario, or at least with some funding to purchase him a modern computer and Matlab licence to more easily fix things.
> > > >
> > > > Ian
> > > >
> > > >
> > >
> >
>